The present invention relates to ion implantation in semiconductor fabrication, and more particularly to evaluation of the implantation dosages of molecular nitrogen on semiconductor wafers.
Ion implantation is commonly used for introducing foreign material into semiconductors during the manufacture of integrated circuits. The foreign atoms that are inserted into the semiconductor are dopants such as arsenic, phosphorus, and boron. The implantation dosage of those atoms is usually measured by the resistance on the wafers or by a measurement tool, such as Therma-Probe(copyright) 500 (TP500)(Therma-Wave, Inc.).
However, the resistance on the wafers after molecular nitrogen (N2) implantation can be too high to be measured. Therefore, the tool, such as TP500 from Therma-Wave, Inc., is used for the measurement of the molecular nitrogen implantation concentration. However, the precision of the TP500 is not satisfied because of inherent bias.
FIG. 1 shows the relation curve between the nitrogen implantation dosage and the measured TW signal values. The TW signal values are obtained by the TP500 implantation measurement tool. Conventionally, there is inherently 3-6% deviation in measurement tool TP500. From the results shown in FIG. 1, the dosage variation is 1E12/per-TW, which means that the linear relation between the molecular nitrogen implantation dosage and the measured TW signal values are not ideal for monitoring the implantation dosage of molecular nitrogen.
Another evaluation of molecular nitrogen implantation is shown in FIG. 2. The curve in FIG. 2 represents TW signal values measured after rapid Thermal process (RTP) in various implantation voltages. The rapid Thermal process (RTP) is conducted at 1100xc2x0 C. and the implantation dosage is 2E14/cm3. The TW signal values are still obtained by the TP500 implantation measurement tool. The relation between measured TW signal values and implantation voltages are more linearly correlated than that between the measured TW signal values and implantation dosages in FIG. 1. However, the curve in FIG. 2 is not suitable for monitoring the molecular nitrogen implantation because the actual implantation concentration will vary in different implantation voltages.
It has been one object of the present invention to provide a method to monitor the molecular nitrogen implantation dosage more precisely and non-destructively.
Another object of the present invention is to build the relation between molecular nitrogen implantation dosage and the thickness of oxide layer after thermal process.
A still further object of the present invention is to monitor the molecular nitrogen implantation using the relation between nitrogen implantation dosage and the thickness of oxide layer after thermal process.
An additional object of the present invention is to monitor the stability of the implantation tool using the relation between molecular nitrogen implantation dosage and the thickness of oxide layer after thermal process.
To achieve the above-mentioned objects, a relation between the molecular nitrogen implantation dosage and the thickness of oxide layer is built. The semiconductor wafers are first implanted with various concentrations of molecular nitrogen. After implantation, the implanted wafers and a non-implanted wafer are subjected to thermal process to grow oxide layer. The thickness of oxide layer on the wafers with various implantation dosage is measured. The implanted nitrogen on the wafers suppresses the growth of oxide layer. Therefore, the higher implantation dosage, the thinner the oxide layer on the wafers. A suppression ratio is computed from the difference in thickness of the oxide layer between the implanted and non-implanted semiconductor wafers to express the thickness variation. Then, a relation between the suppression ratio and the dosages of molecular nitrogen is built.
In one aspect of the present invention, the relation between the suppression ratio of oxide thickness and the dosages of molecular nitrogen can be used to predict the molecular nitrogen dosage needed to grow a predetermined thickness of oxide. A molecular nitrogen dosage needed to grow a predetermined thickness of oxide layer on a process wafer is estimated by inputting the predetermined thickness into the relation.
In another aspect of the present invention, the relation between the suppression ratio and the dosages of molecular nitrogen can also be used to assess the stability of an ion implanter. A monitor wafer is implanted with a predetermined molecular nitrogen dosage and then subjected to thermal process to grow an oxide layer on a surface of the monitor wafer. An estimative molecular nitrogen dosage can be computed by inputting the thickness of oxide layer on the monitor wafer into the relation. By evaluating the deviation between the estimative molecular nitrogen dosage and the predetermined molecular nitrogen dosage of the monitor wafer, the ion implanter is unstable if the deviation is larger than a specific range.